KR20230003051A - Image alignment for noisy images - Google Patents

Abstract

A method and system for aligning an image of a specimen is provided. One method includes reducing noise in a test image generated for a specimen by an imaging subsystem to produce a denoised test image. The method also includes detecting one or more patterned features in the denoised test image extending in at least a horizontal or vertical direction. The method also includes designating a region of the denoised test image in which the detected one or more patterned features are located as a region of interest in the denoised test image. The method further includes aligning the denoised test image to a reference image of the specimen using only the region of interest in the denoised test image and a corresponding region in the reference image.

Description

Image alignment for noisy images

The present invention relates generally to a method and system for aligning images of a specimen, including images of particularly noisy specimens.

The following descriptions and examples are not admitted to be prior art by their inclusion in this section.

For example, manufacturing semiconductor devices, such as logic devices and memory devices, typically uses a number of semiconductor fabrication processes to process a substrate, such as a semiconductor wafer, to form various features and multiple levels of the semiconductor device. includes doing For example, lithography is a semiconductor manufacturing process that involves transferring a pattern from a reticle to a resist disposed on a semiconductor wafer. Additional examples of semiconductor fabrication processes include, but are not limited to, chemical-mechanical polishing (CMP), etching, deposition, and ion implantation. Multiple semiconductor devices can be fabricated in an arrangement on a single semiconductor wafer and then separated into individual semiconductor devices.

Inspection processes can be used at various stages during the semiconductor fabrication process to detect defects on wafers, facilitating higher yields and higher profits in the fabrication process. Inspection has always been an important part of manufacturing semiconductor devices such as ICs, for example. However, as the dimensions of semiconductor devices decrease, inspection becomes even more important for the successful manufacture of acceptable semiconductor devices, since smaller defects can cause the device to fail.

Defect review usually entails redetecting defects that have been detected as defects by the inspection process and generating additional information about the defect at higher resolution using a high-magnification optical system or scanning electron microscope (SEM) . Thus, defect review is performed at discrete locations on the wafer where defects are detected by inspection. Higher resolution data on defects generated by defect review are better suited for determining the defect's properties, such as profile, roughness, and more accurate size information.

Metrology processes are also used at various stages during the semiconductor manufacturing process to monitor and control this process. A metrology process differs from an inspection process in that a metrology process is used to measure one or more characteristics of a wafer that cannot be determined using currently used inspection tools, unlike an inspection process in which defects are detected on the wafer. For example, a metrology process may be used to measure one or more characteristics of a wafer, such as, for example, the dimensions of features formed on the wafer during the process (eg, line width, thickness, etc.), so that the performance of the process may be measured in one or more characteristics. can be determined from Further, if one or more characteristics of the wafer are unacceptable (e.g., outside a predetermined range for the characteristic(s)), then the measurement of one or more characteristics of the wafer determines whether additional wafers made by the process have the acceptable characteristic(s). can be used to change one or more parameters of a process to

The metrology process also differs from the defect review process in that the metrology process can be performed where no defects have been detected, unlike the defect review process in which defects detected by inspection are re-visited in the defect review. . That is, unlike defect review, the location at which a metrology process is performed on a wafer may be independent of the outcome of an inspection process performed on the wafer. In particular, the location where the metrology process is performed can be selected regardless of the inspection result. Also, unlike defect review, where the location on the wafer at which the defect review is to be performed cannot be determined until the inspection results for the wafer are generated and available, since the location on the specimen at which metrology is performed can be selected independently of the inspection result. , the location where the metrology process is performed can be determined before the inspection process is performed on the wafer.

One of the challenges of a quality control type process like the one described above is to align one image to another with sufficient accuracy. Image alignment is typically performed to align a test image to a reference image so that a difference between the test image and the reference image can be determined. These differences can then be used to detect defects in the case of inspection and defect review and to determine relative measurements in the case of metrology. Clearly, therefore, if the test image and the reference image are not precisely aligned, misalignment can lead to erroneous results in the results produced by this process.

Some methods currently used for image alignment in semiconductor quality control type processes take two raw images and align them directly to each other with normalized cross correlation (NCC). NCC is a statistics-based method for calculating the correlation between two samples. The simplest form of NCC is the cosine of the angle between two vectors a and b.

This method is effective and practical when the templates, i.e., the reference image and the test image, have obvious common features and relatively little noise.

However, currently used methods and systems for image alignment have several disadvantages. For example, currently used methods and systems do not consider the effect of noise on vertical or horizontal features of an image. Also, in some images there are only a few horizontal or vertical features, and the alignment position can be misled by noise.

Accordingly, it would be advantageous to develop a system and method for aligning an image of a specimen that does not have one or more of the disadvantages described above.

The following description of various embodiments should in no way be construed as limiting the subject matter of the appended claims.

One embodiment relates to a system configured for aligning an image of a specimen. The system includes an imaging subsystem configured to generate images of the specimen. The system also includes one or more computer subsystems configured to reduce noise in a test image generated of the specimen by the imaging subsystem to produce a denoised test image. The one or more computer subsystems are also configured to detect one or more patterned features in the denoised test image extending in at least a horizontal or vertical direction. Additionally, one or more computer subsystems are configured to designate a region of the denoised test image in which the detected one or more patterned features are located as a region of interest (ROI) in the denoised test image. The one or more computer subsystems are also configured to align the denoised test image to the reference image for the specimen using only the ROI in the denoised test image and the corresponding region in the reference image. This system may also be configured as described herein.

Another embodiment relates to a computer-implemented method for aligning an image of a specimen. The method includes the steps described above for reducing noise, detecting one or more patterned features, designating areas, and aligning. Method steps are performed by one or more computer subsystems coupled to the imaging subsystem configured as described above.

Each of the steps of the foregoing method may be performed as further described in this disclosure. Each of the embodiments of the method described above may include any other step(s) of any other method(s) described herein. The method may be performed by any of the systems described herein.

Another embodiment relates to a non-transitory computer readable medium storing program instructions executable on a computer system to perform a computer implemented method for aligning an image of a specimen. The computer-implemented method includes the steps of the method described above. Computer readable media may also be configured as described herein. Steps of the computer-implemented method may be performed as also described herein. Moreover, the method by which the program instructions are implemented in an executable computer may include any other step(s) of any other method(s) described herein.

Further advantages of the present invention will become apparent to those skilled in the art having the benefit of the following detailed description of preferred embodiments with reference to the accompanying drawings.
1 and 2 are schematic diagrams illustrating side views of an embodiment of a system constructed as described herein;
3 includes examples of test images before and after denoising.
4-6 are flow diagrams illustrating embodiments of steps that may be performed by one or more computer subsystems described herein to align an image of a specimen; And
7 is a block diagram illustrating one embodiment of a non-transitory computer readable medium storing program instructions for causing a computer system to perform the computer implemented methods described herein.
Although this invention is capable of many modifications and alternative forms, specific embodiments thereof are illustrated by way of example in the drawings and described in detail herein. The drawings may not be drawn to scale. However, the drawings and their detailed description are not intended to limit the invention to the particular forms disclosed; on the contrary, the intent is to all modifications, equivalents and alternatives within the spirit and scope of the invention as defined by the appended claims. It should be understood that it is intended to cover (cover).

Referring now to the drawings, it is noted that the drawings are not drawn to scale. In particular, the scale of some of the elements in the drawings is greatly exaggerated to emphasize the characteristics of the elements. It is also noted that the drawings are not drawn to scale. Elements shown in more than one figure that may be similarly constructed are indicated using the same reference numerals. Unless otherwise indicated herein, any element described and illustrated may include any suitable commercially available element.

One embodiment relates to a system configured for aligning an image of a specimen. Some embodiments relate to improving alignment of images, eg, scanning electron microscope (SEM) images, using denoising and region-of-interest (ROI) designation. One of the major challenges in the image processing community is image alignment, especially when the goal is to align two images within sub-pixel accuracy. The problem becomes even more difficult when there are images, for example, SEM images with repetitive patterned features that lack any other distinguishable structures (or have limited distinguishable structures). The embodiments described herein introduce new methods and systems for improving image alignment to reach sub-pixel accuracy.

Embodiments described herein may be used to improve the accuracy of an image alignment process, such as, for example, between a test image and a reference image including SEM test and reference images. As mentioned above, often SEM images are relatively noisy and are preferably aligned within pixel accuracy. One major problem with image alignment stems from the fact that if the image has artifacts (i.e., noise), for example, features such as edges may be jagged and parallel lines may be distorted. Also, some images have only a small number of horizontal or vertical features, and these features will have less weight than noise has in an image alignment process, for example normalized cross correlation (NCC). can As a result, the likelihood of misalignment increases due to the roughness of the patterned feature edges in the image.

As further described herein, embodiments provide a new approach for improved image alignment. One approach involves denoising the image using a structure-based denoising method. Another approach is to use singular value decomposition (SVD) to preserve only the main features of the image to achieve the desired result. Combining these two approaches can even further improve alignment if your computational budget allows. In addition, as described further herein, an embodiment detects a horizontal feature or a vertical feature in an image, sets this aperture as a region of interest (ROI), and sets the ROI as an input to an alignment method, e.g., NCC. and one of the denoising approaches described here.

In one embodiment, the specimen is a wafer. The wafer may include any wafer known in the semiconductor art. In another embodiment, the specimen is a reticle. The reticle may include any reticle known in the semiconductor art. Although some embodiments may be described herein with respect to a wafer or wafers, the embodiments are not limited to the specimens in which they may be used. For example, embodiments described herein may be used with specimens such as, for example, reticles, flat panels, personal computer (PC) boards, and other semiconductor specimens.

One embodiment of such a system is shown in FIG. 1 . In some embodiments, the system includes an imaging subsystem configured to generate images of the specimen. The imaging subsystem may include at least an energy source and a detector. The energy source is configured to generate energy directed to the specimen. The detector is configured to detect energy from the specimen and generate an output in response to the detected energy.

In one embodiment, the imaging subsystem is a light-based imaging subsystem. For example, in the embodiment of the system shown in FIG. 1 , imaging subsystem 10 includes an illumination subsystem configured to direct light onto specimen 14 . The lighting subsystem includes at least one light source. For example, as shown in FIG. 1 , the lighting subsystem includes a light source 16 . In one embodiment, the illumination subsystem is configured to direct light onto the specimen at one or more angles of incidence, which may include one or more oblique angles and/or one or more vertical angles. For example, as shown in FIG. 1 , light from light source 16 is directed through optical element 18 and then through lens 20 to beam splitter 21, where beam splitter 21 transmits light. is directed at the specimen 14 at a normal angle of incidence. The angle of incidence may include any suitable angle of incidence, which may vary depending on, for example, the nature of the specimen and the process to be performed on the specimen.

The illumination subsystem can be configured to direct light to the specimen at different times and at different angles of incidence. For example, the imaging subsystem can be configured to change one or more properties of one or more elements of the illumination subsystem such that light can be directed to the specimen at a different angle of incidence than shown in FIG. 1 . In one such example, the imaging subsystem may be configured to move light source 16, optical element 18, and lens 20 to direct light at different angles of incidence to the specimen.

Optionally, the imaging subsystem can be configured to direct light onto the specimen at more than one angle of incidence at the same time. For example, an imaging subsystem may include more than one illumination channel, and one of the illumination channels may include light source 16, optical element 18, and lens 20 as shown in FIG. and another one of the illumination channels (not shown) may include similar elements that may be configured differently or identically, or at least a light source and possibly as described further herein, for example. may contain one or more other components, such as If this light is directed to the specimen at the same time as other light, one or more properties (eg, wavelength, polarization, etc.) of the light directed to the specimen at different angles of incidence will cause the light resulting from illumination of the specimen at the different angles of incidence to be detected by the detector(s). may be different so that they can be distinguished from each other.

In another example, the lighting subsystem may include only one light source (eg, source 16 shown in FIG. 1 ), and the light from the light source may be directed to one or more optical elements (not shown) of the lighting subsystem. (e.g., based on wavelength, polarization, etc.) into different optical paths. Light from each of the different optical paths can then be directed to the specimen. Multiple illumination channels may be configured to direct light to the specimen at the same time or at different times (eg, when different illumination channels are used to sequentially illuminate the specimen). In another example, the same illumination channel can be configured to direct light with different characteristics to the specimen at different times. For example, in some cases, optical element 18 may be configured as a spectral filter, the properties of which may be configured in a variety of different ways (e.g., spectral by replacing the filter). The illumination subsystem may have any other suitable configuration known in the art for directing light having different or identical characteristics to the specimen sequentially or simultaneously at different or equal angles of incidence.

In one embodiment, light source 16 may include a broadband plasma (BBP) light source. In this way, the light generated by the light source and directed to the specimen may include broadband light. However, the light source can include any suitable laser known in the art and can be configured to produce light at any suitable wavelength(s) known in the art, for example any other suitable laser. A light source may be included. Also, the laser can be configured to produce monochromatic or nearly monochromatic light. In this way, the laser may be a narrowband laser. The light source may also include a polychromatic light source that produces light at multiple discrete wavelengths or wavelength bands.

Light from optical element 18 may be focused by lens 20 to beam splitter 21 . Although lens 20 is shown in FIG. 1 as a single refractive optical element, in practice lens 20 may include multiple refractive and/or reflective optical elements that combine to focus light from the optical elements onto a specimen. . The illumination subsystem shown in FIG. 1 and described herein may include any other suitable optical element (not shown). Examples of such optical elements include polarization component(s), spectral filter(s), spatial filter(s), reflective optical element(s), apodizers, which may include any such suitable optical elements known in the art. (s), beam splitter(s), aperture(s), etc., but are not limited thereto. Additionally, the system may be configured to change one or more elements of the illumination subsystem based on the type of illumination used for imaging.

The imaging subsystem may also include a scanning subsystem configured such that light is scanned over the specimen. For example, the imaging subsystem may include stage 22 on which specimen 14 is placed during imaging. The scanning subsystem may include any suitable mechanical and/or robotic assembly (including stage 22) that may be configured to move the specimen such that light may be scanned over the specimen. Additionally or alternatively, the imaging subsystem may be configured such that one or more optical elements of the imaging subsystem perform some scanning of light over the specimen. The light can be scanned over the specimen in any suitable manner.

The imaging subsystem further includes one or more detection channels. At least one of the one or more detection channels includes a detector configured to detect light from the specimen due to illumination of the specimen by the imaging subsystem and to generate an output in response to the detected light. For example, the imaging subsystem shown in FIG. 1 has one channel formed by collector 24, element 26 and detector 28 and another channel formed by collector 30, element 32 and detector 28. (34). As shown in Figure 1, the two detection channels are configured to collect and detect light at different collection angles. In some examples, one detection channel is configured to detect specularly reflected light and the other detection channel is configured to detect light that is not specularly reflected (eg, scattered, diffracted, etc.) from the specimen. However, two or more detection channels may be configured to detect the same type of light (eg, specularly reflected light) from the specimen. 1 shows an embodiment of an imaging subsystem that includes two detection channels, the imaging subsystem may include a different number of detection channels (eg, just one detection channel or two or more detection channels). . Although each of the collectors is shown in FIG. 1 as a single refractive optical element, each of the collectors may include one or more refractive optical element(s) and/or one or more reflective optical element(s).

The one or more detection channels may be any known in the art, such as, for example, photo-multiplier tubes (PMT), charge coupled devices (CCD), time delay integration (TDI) cameras. of suitable detectors. The detector may also include a non-imaging detector or an imaging detector. Where the detectors are non-imaging detectors, each detector may be configured to detect certain characteristics of the scattered light, such as, for example, intensity, but not configured to detect characteristics, such as, for example, as a function of position within the imaging plane. may not be As such, the output produced by each detector included in each detection channel may be a signal or data, but may not be an image signal or image data. In such a case, a computer subsystem, such as, for example, computer subsystem 36 of the system, may be configured to generate an image of the specimen from the non-imaging output of the detector. In other cases, however, the detector may be configured as an imaging detector configured to generate an imaging signal or image data. Accordingly, the system can be configured to generate images in a number of ways.

1 is provided herein to generally illustrate the configuration of an imaging subsystem that may be included in system embodiments described herein. Obviously, when designing a commercial imaging system, the imaging subsystem configurations described herein may be altered to optimize the system's performance as it normally performs. Additionally, the systems described herein can be used with existing imaging systems, such as, for example, the 29xx and 39xx series of tools commercially available from KLA (e.g., by adding the functionality described herein to an existing inspection system). by doing) can be implemented. For some such systems, the embodiments described herein may be provided as optional functionality of the imaging system (eg, in addition to other functionality of the imaging system). Alternatively, the imaging subsystem described herein can be designed “from scratch” to provide a completely novel imaging system.

Computer subsystem 36 of the system may include in any suitable manner (e.g., "wired" and/or "wireless" transmission media) such that the computer subsystem can receive output generated by the detector during scanning of the specimen. may be coupled to a detector of the imaging subsystem (via one or more transmission media that may be capable of). Computer subsystem 36 may be configured to perform a number of functions using the output of the detector as described herein and any other functions described further herein. This computer subsystem may be further configured as described herein.

This computer subsystem (and other computer subsystems described herein) may also be referred to herein as computer system(s). Each of the computer subsystem(s) or system(s) described herein may take a variety of forms including a personal computer system, image computer, mainframe computer system, workstation, network appliance, Internet appliance, or other device. can In general, the term "computer system" can be broadly defined to include any device having one or more processors that execute instructions from a memory medium. The computer subsystem(s) or system(s) may also include any suitable processor known in the art, such as, for example, a parallel processor. Additionally, the computer subsystem(s) or system(s) may include a computer platform with high-speed processing and software as a stand-alone or networked tool.

If the system includes more than one subsystem, the different computer subsystems can be coupled together so that images, data, information, instructions, etc. can be transferred between the computer subsystems. For example, computer subsystem 36 may communicate with computer subsystems 102 (in FIG. 1) by any suitable transmission medium, which may include any suitable wired and/or wireless transmission media known in the art. as shown by dotted lines). Two or more of these computer subsystems may also be effectively coupled by a shared computer readable storage medium (not shown).

Although the imaging subsystem is described above as being an optical or light-based subsystem, in another embodiment, the imaging subsystem is an electronic-based imaging subsystem. For example, in one embodiment, the energy directed to the specimen includes electrons and the energy detected from the specimen includes electrons. In this way, the energy source may be an electron beam source. In one such embodiment shown in FIG. 2 , the imaging subsystem includes an electron column 122 coupled to a computer subsystem 124 .

As also shown in FIG. 2 , the electron column includes an electron beam source 126 configured to generate electrons that are focused on a specimen 128 by one or more elements 130 . The electron beam source may include, for example, a cathode source or an emitter tip, and one or more elements 130 may include, for example, a gun lens, an anode, a beam limiting aperture ), a gate valve, a beam current selection aperture, an objective lens, and a scanning subsystem, all of which may include any such suitable elements known in the art.

Electrons (eg, secondary electrons) returned from the specimen may be focused to detector 134 by one or more elements 132 . One or more elements 132 may include a scanning subsystem, which may be, for example, the same scanning subsystem included within element(s) 130 .

The electron column may include any other suitable element known in the art. Additionally, electronic columns are disclosed in US Pat. No. 8,664,594 to Jiang et al., issued Apr. 4, 2014, and U.S. Patent No. 8,692,204 to Kojima et al., issued Apr. 8, 2014, Apr. 2014 may also be configured as described in U.S. Patent No. 8,698,093 to Gubbens et al, issued May 15, and U.S. Patent No. 8,716,662 to MacDonald et al, issued May 6, 2014 However, these US patents are incorporated by reference as if fully set forth herein.

Although the electron column is shown in Figure 2 as being configured such that electrons are directed at the specimen at an oblique angle of incidence and scattered from the specimen at another oblique angle, it is understood that the electron beam may be directed at and scattered from the specimen at any suitable angle. It should be. Additionally, the e-beam subsystem can be configured to use multiple modes to generate images of the specimen (eg, at different illumination angles, collection angles, etc.). The multiple modes of the e-beam subsystem may differ in any image generation parameter(s) of the subsystem.

Computer subsystem 124 may be coupled to detector 134 as described above. The detector can detect electrons that return from the surface of the specimen and thereby form an electron beam image of the specimen. The electron beam image may include any suitable electron beam image. Computer subsystem 124 may be configured to perform any of the functions described herein using the output of the detector and/or the electron beam image. Computer subsystem 124 may be configured to perform any additional step(s) described herein. A system that includes the imaging subsystem shown in FIG. 2 may be further configured as described herein.

2 is provided in this disclosure to generally illustrate the configuration of an electronics-based imaging subsystem that may be included in an embodiment described herein. As with the optical subsystems described above, the e-beam subsystem configurations described in this disclosure can be modified to optimize the performance of the subsystems, which is commonly done when designing commercial imaging systems. In addition, the systems described herein may be implemented using existing imaging systems (eg, by adding the functionality described herein to existing systems). For some such systems, the embodiments described herein may be provided as optional functionality of the system (eg, in addition to other functionality of the system). Alternatively, the system described herein may be designed “from scratch” to provide a completely novel system.

Although the imaging subsystem is described above as being a light based or electron beam based subsystem, the imaging subsystem may be an ion beam based subsystem. This imaging subsystem may be configured as shown in FIG. 2 except that the electron beam source may be replaced with any suitable ion beam source known in the art. Thus, in one embodiment, the energy directed at the specimen includes ions. In addition, imaging subsystems include, for example, commercially available focused ion beam (FIB) systems, helium ion microscopy (HIM) systems, and secondary ion mass spectroscopy. ; SIMS) system.

An imaging subsystem described herein may be configured to generate an output, eg, an image, of a specimen having multiple modes. A “mode” is usually defined by the value of a parameter of the imaging subsystem used to generate an image of the specimen (or output used to generate an image of the specimen). Thus, the modes may have different values for at least one of the parameters of the imaging subsystem (except for the location on the specimen at which the output is generated). For example, in an optical subsystem, different modes may use different wavelength(s) of light for illumination. The modes may differ in illumination wavelength (eg, by using different light sources, different spectral filters, etc. for different modes) as described further herein. In another example, different modes may use different illumination channels of the optical subsystem. For example, as noted above, an optical subsystem may include more than one illumination channel. As such, different illumination channels can be used for different modes. The modes may also or alternatively differ in one or more collection/detection parameters of the optical subsystem. Modes differ in one or more changeable parameters of the imaging subsystem (e.g., illumination polarization(s), angle(s), wavelength(s), etc., detection polarization(s), angle(s), wavelength(s), etc.) can do. The imaging subsystem may be configured to scan the specimen in different modes, either in the same scan or in different scans, for example with the ability to use multiple modes to scan the specimen simultaneously.

In a similar manner, the output produced by the e-beam subsystem may include an output, eg, an image, produced by the e-beam subsystem using two or more different values of a parameter of the e-beam subsystem. A number of modes of the e-beam subsystem may be defined by values of parameters of the e-beam subsystem used to generate outputs and/or images for the specimen. Thus, the modes may have different values for at least one of the e-beam parameters of the e-beam subsystem. For example, different modes may use different angles of incidence for illumination.

The subsystems described in this disclosure and shown in FIGS. 1 and 2 can be modified in one or more parameters to provide different imaging capabilities depending on the application to be used. In one such example, the imaging subsystem shown in FIG. 1 may be configured to have a higher resolution when used for defect review or metrology than for inspection. That is, the embodiment of the imaging subsystem shown in FIGS. 1 and 2 is an imaging subsystem that can be tailored in a number of ways that will be apparent to those skilled in the art to produce imaging subsystems with different output generating capabilities that are more or less suitable for different applications. Some common and varied configurations for

As mentioned above, the optical subsystem, electronic subsystem, and ion beam subsystem are configured to scan energy (eg, light, electrons, etc.) over the physical version of the specimen to create an image of the physical version of the specimen. In this way, the optical subsystem, electronic subsystem, and ion beam subsystem can be configured as “real” subsystems rather than “virtual” subsystems. However, the storage media (not shown) and computer subsystem(s) 102 illustrated in FIG. 1 may be configured as “virtual” systems. In particular, the storage medium and computer subsystem(s) are described in commonly assigned U.S. Patent Nos. 8,126,255 to Bhaskar et al. (Duffy) et al., 9,222,895, all of which are incorporated herein by reference as if fully set forth. The embodiments described herein may be further configured as described in these patents.

One or more computer subsystems are configured to reduce noise in test images generated of the specimen by the imaging subsystem to produce denoised test images. 3 shows examples of images before and after denoising. In particular, image 300 is the original SEM test image and image 302 is a denoised version of that SEM test image. As can be seen by comparing image 300 and image 302, the edges of the patterned features in the denoising image are much smoother after denoising. Further, denoising reduced noise in unpatterned portions of the image (eg, relatively large spaces between patterned features) as well as made the image brighter and increased the contrast between features in the image and the background. All of these changes will make image alignment easier and more accurate as described herein. While the image of FIG. 3 illustrates a specific number, arrangement, type, orientation, shape, etc. of patterned features, the embodiments described herein are not limited to the type of image, patterned features in image, type of specimen, etc. Examples can be used for That is, the embodiments described herein are not limited to images and specimens, and the embodiments may be used for these.

In some embodiments, the test image includes repetitive patterned features. For example, as shown in FIG. 3 , some of the patterned features of the original test image 300 are repeated within the image. As explained above, image alignment can be more difficult when there are repetitive features without any other (or limited) distinguishable features. For the test image example shown in FIG. 3, there is at least one patterned feature in the test image that is distinguishable from all other patterned features in the image. However, some images that must be aligned with each other for the processes described herein are not guaranteed to have even one feature distinguishable from other patterned features in the image. However, embodiments described herein may be used to align images with few or even no unique patterned features relative to other patterned features in the image.

In one embodiment, reducing noise includes structure-based denoising. As used herein, the term "structure-based denoising" is a generic term used to refer to image denoising that is based on and preserves the structure seen in an image. One embodiment of a suitable structure-based denoising method that can be used to reduce noise in images described herein is the block-matching and 3D filtering (BM3D) algorithm. In general, the BM3D algorithm groups image fragments based on similarity. The image fragments do not have to be separate, but they must be the same size. Whether the fragments are grouped can be determined by applying a threshold to the degree of dissimilarity to the reference fragment. This grouping technique is commonly referred to as block matching. However, BM3D can group macroblocks within a single frame, and then stack all the image pieces in the group to form a 3D cylinder-like shape. BM3D may then include filtering performed on all fragment groups, followed by a linear transformation. You can then perform transformation domain reduction steps, eg Wiener filtering, and then recreate all (filtered) slices by inverting the linear transformation. The image is then converted back to 2D form. In addition, by applying a weighted average to all overlapping image segments, it is possible to maintain a distinct signal while filtering out noise. Structure-based denoising is described in Cheng et al. in "Image denoising algorithm based on structure and texture part," 2016 12th International Conference on Computational Intelligence and Security, December 2016, IEEE, pp. 147-151, which is incorporated herein by reference as if fully set forth. Embodiments described herein may be further configured as described in this reference.

In another embodiment, reducing noise includes singular value decomposition. For example, singular value decomposition (SVD) can be used to pre-process images, such as SEM images, thus improving the alignment success rate. SVD decomposes a matrix into orthogonal components from which optimal sub-rank approximations can be obtained. The largest object component in an image found using SVD generally corresponds to the eigen-images associated with the largest singular value, whereas the image noise corresponds to the eigen-image associated with the smallest singular value. Thus, SVD can be a particularly useful tool for isolating noise from patterned feature images in the embodiments described herein.

In some embodiments, reducing noise comprises selecting a SVD, a predetermined number of eigenvectors having a maximum value in one or more matrices output by the SVD, and reconstructing the test image from the selected eigenvectors to obtain a denoised test image. including generating In one such embodiment, as shown in FIG. 6 , test image 600 and reference image 602 may be input to separate SVD stages 604 and 606 , respectively. In this step, each input image can be decomposed into a left matrix, a right matrix, and a diagonal matrix. The output of each SVD step can be input to a small eigenvector removal step. For example, the output of SVD stage 604 can be input to small eigenvector removal stage 608 and the output of SVD stage 606 can be input to small eigenvector removal stage 610 . In the small eigenvector removal step, only a few (e.g., 10) of the largest eigenvectors (EgV) can be kept and others can be removed. In this way, relatively small unwanted features that can negatively affect the alignment of the image can be removed.

One or more computer subsystems can then reconstruct an image containing only the relatively large main structures and then feed it to the main alignment method. For example, as shown in FIG. 6 , the output of the small eigenvector removal step 608 can be input to the image reconstruction and ROI step 612 and the output of the small eigenvector removal step 610 can be used to image reconstruction. and ROI step 614. Image reconstruction may be performed as described above and ROI identification and designation may be performed as further described herein. The output of the image reconstruction and ROI steps may be input to an alignment step 616, which may be performed as further described herein. In this way, embodiments described herein may use SVD to remove edge roughness and noise from an image, eg, a SEM image, before applying an alignment method or algorithm to the image.

In a further embodiment, reducing noise includes structure-based denoising followed by singular value decomposition. For example, the input images shown in FIG. 6 , namely test image 600 and reference image 602 , have been previously denoised by structure-based denoising performed according to any of the embodiments described herein. may be an image. In this way, if the computational budget permits, structure-based denoising can be combined with SVD to improve alignment more than the denoising method alone provides.

The one or more computer subsystems are also configured to detect one or more patterned features in the denoised test image extending in at least a horizontal or vertical direction. As used herein, “extending in at least a horizontal or vertical direction” is intended to mean that the patterned feature(s) have some extended lateral dimension in one of these directions. For example, a line or trench on a specimen such as a wafer is considered a patterned feature extending in at least one direction, whereas a contact hole or other circular or square structure is considered a patterned feature extending in at least a horizontal or vertical direction. not considered In this way, generally a "patterned feature extending in at least the horizontal or vertical direction" is a patterned feature having one dimension in the x or y direction of the specimen greater than another dimension of the patterned feature extending in the opposite direction. It can be a feature. A "patterned feature extending in at least a horizontal or vertical direction" also means that the entirety of the patterned feature does not extend in at least a horizontal or vertical direction (e.g., not all portions of the patterned feature extend in the same direction) and/or or patterned features that extend in both horizontal and vertical directions (eg, where one portion of the patterned features extends in the horizontal direction and another portion of the patterned features extends in the vertical direction).

Some examples of different patterned features extending in at least the horizontal or vertical direction are shown in region 304 of denoised test image 302 of FIG. 3 . In particular, the patterned feature 304a extends in a vertical direction rather than a horizontal direction. The patterned features 304b may or may not be considered to extend in the vertical direction and extend in the horizontal direction. Patterned feature 304c extends only in the vertical direction and patterned feature 304d extends vertically, then horizontally, and again vertically when traversing across the patterned feature. All of these patterned features may be determined to extend in at least a horizontal or vertical direction by embodiments described herein.

Patterned features extending in at least a horizontal direction (i.e., “horizontal patterned features”) or patterned features extending in a vertical direction (i.e., “vertical patterned features”) are, for example, gray-level based thresholding ), image projection, and the like. Horizontally and/or vertically patterned features may be detected in the denoised test image using the denoised test image itself, for example by inputting the denoised test image to a detection step. Detecting horizontal patterned features and/or vertical patterned features can be important in identifying suitable features for use in the alignment steps described herein. Whether horizontal features and/or vertical features are detected and used for the alignment described herein may depend on the specimen, what patterned features are formed in the imaging area on the specimen, and possibly the image alignment algorithm (e.g., horizontal patterning If there are both patterned features and vertically patterned features, which feature(s) are detected and used for alignment may depend on the type of features available to the alignment algorithm; the alignment algorithm is not both horizontal and vertical, but only if it performs better on patterned features that extend in only one of these directions; and so on).

The one or more computer subsystems are also configured to designate a region of the denoised test image in which the detected one or more patterned features are located as a region of interest (ROI) in the denoised test image. The ROI in the denoised test image may contain only a single contiguous region in the denoised test image. For example, in the denoised test image shown in FIG. 3, region 304 that includes at least horizontally or vertically extending patterned features 304a, 304b, 304c, and 304d is one or more computer subs. It can be designated as an ROI by the system. Alternatively, an ROI may include more than one region, some of which are distinct or mutually exclusive. For example, horizontal patterned features and/or vertical patterned features may be found in the denoised test image at various locations depending on the design for the specimen within the imaged area. Thus, for some imaged areas, e.g., not including patterned features or areas of the denoised test image that are not useful (or even detrimental) for alignment, there are horizontal patterned features and/or vertical patterned features. It may not be possible or advantageous to specify a single contiguous region of the denoised test image in which all (or a sufficient number) are located.

Then, in some such cases, a designated ROI may be defined as an area containing fewer than both detected horizontal patterned features and/or vertical patterned features. In other such examples, more than one designated ROI may be defined, and each of the designated ROIs may include a different subset of horizontal patterned features and/or vertical patterned features. Moreover, in some cases, there may be patterned features that do not extend in a horizontal or vertical direction adjacent to other patterned features that extend in a horizontal or vertical direction. In such cases, where computationally meaningful, such non-horizontal patterned features and/or non-vertical patterned features may be included in the ROI along with the horizontal patterned features and/or vertical patterned features. Thus, in general, the embodiments described herein may designate an ROI as any region in a denoised test image that includes at least one horizontally patterned feature and/or vertically patterned feature, in the denoised test image. Not all of the detected horizontal patterned features and/or vertical patterned features may be included in the ROI, and the ROI may include some patterned features that are not horizontal patterned features and/or vertical patterned features.

The one or more computer subsystems are also configured to align the denoised test image to the reference image for the specimen using only the ROI in the denoised test image and the corresponding region in the reference image. By aligning the denoised test image to the reference image, the computer subsystem(s) effectively align the original test image to the reference image. Specifying the ROI and using only the ROI for the alignment step can be critical to achieving the desired accuracy of image alignment. For example, the denoising methods described herein may not remove all noise from all test images that may be selected for image alignment. Using only the ROI for image alignment can improve the accuracy of image alignment by reducing the influence of the rest of the noise.

In some embodiments, aligning the denoised test image to the reference image aligns the test image to the reference image with sub-pixel accuracy. As used herein, a “subpixel” is generally defined as being smaller than a pixel of output produced by an imaging subsystem. In this way, as used herein, the term "subpixel accuracy" generally refers to something that has an error smaller than the size (distance from one side to the other) of a single pixel of an image acquired by the imaging subsystem (e.g., image alignment). ) can be defined as determining Performing the alignment using only the denoising step(s), the designated ROI, and the designated ROI described herein allows image alignment to be performed with sub-pixel accuracy. In other words, the denoising step(s), ROI designation step, and alignment step can be performed as described herein to align test and reference images with sub-pixel accuracy.

In a further embodiment, aligning the denoised test image to the reference image includes normalized cross-correlation. NCC may be performed as further described herein or in any other manner known in the art. Although NCC may be generally used by some of the systems and methods described herein to perform the actual alignment step, the alignment described herein may be performed using any other correlation type algorithm or method known in the art. there is. In other words, while NCC is one method that is particularly well suited for aligning the test image and reference image described here, the denoised test image and reference image, along with their designated ROIs, are arbitrary for the alignment steps described here. It can be input into an appropriate sorting method or algorithm.

In some embodiments, most of the patterned features extending at least horizontally or vertically in the test image have a smaller weight in the NCC than noise in the test image. For example, embodiments described herein may be useful for aligning images that contain very few patterned features suitable for alignment and/or are noisy, such that the alignment method relies more on noise in the image than patterned feature images. can be particularly advantageous. In particular, by denoising a test image as described herein, designating an ROI in the denoised test image where patterned features suitable for use in alignment are located, and using only the ROI for alignment, The described embodiment makes image alignment fairly accurate (eg, with sub-pixel accuracy) even with significant test image noise and/or few or minimal patterned features suitable for alignment.

The reference images used in the embodiments described herein may be different types of reference images depending on the application in which they are used. Some reference images are created from the specimen being processed. In one embodiment, the reference image is a denoised reference image, the initial reference image is created for the specimen by the imaging subsystem at a location on the specimen that corresponds to the location at which the test image is created, and the one or more computer subsystems and to reduce noise in the initial reference image prior to alignment to produce a denoised reference image. In one such example, test images and reference images may be acquired at different adjacent dies on the specimen and at approximately corresponding locations within those different dies. When both the test image and the reference image are acquired using the specimen, both the test image and the reference image may exhibit relatively similar levels of noise. Thus, if a test image is denoised for alignment purposes as described herein, it is likely to be advantageous to also denoise the reference image for alignment purposes. In this case, both the test image and the reference image can be denoised in the same way using the same parameters (this may be appropriate since the images have relatively similar noise types or levels). This denoising of both the test image and the reference image can be further performed as described herein and shown in FIGS. 4-6 .

In another embodiment, the reference image is not created using the specimen. For example, in some applications described herein, reference images may be created in a manner that does not involve imaging of the specimen. In one such example, the reference image may be a rendered image from a design for a specimen that simulates how the reference image would appear if it were created by imaging the specimen. The design for the specimen may include any suitable design data or information, such as, for example, a graphical data stream (GDS) file. In another example, the reference image can be created from another specimen having the same design printed on it.

In this way, the reference images used in the embodiments described herein may or may not exhibit the same type and level of noise as the test images. For example, a reference image rendered from a design for a specimen may not have the same type and level of noise as the test image, and consequently denoising of such reference image is performed to achieve the image alignment accuracy described herein. may not need to be In other cases, a reference image created from another specimen of the same design may have a relatively similar level and type of noise to the test image, so denoising may be performed on this reference image for alignment purposes.

In some cases, a system or method for generating a reference image may perform denoising of the reference image, making denoising of the reference image unnecessary according to embodiments described herein. For example, rendering the reference image may be performed by another system or method, and the other system or method may denoise the reference image prior to making it usable by embodiments described herein. Thus, the embodiments described herein can use the reference image without any additional denoising.

In some embodiments, the reference image is a denoised reference image, the initial reference image is not created using the specimen, and one or more computer subsystems reduce noise in the initial reference image prior to alignment to obtain the denoised reference image. configured to create For example, embodiments described herein may be configured to generate a reference image without using the physical specimen itself described above (e.g., by rendering a reference image from a design, using another specimen of the same design). by acquiring a reference image, etc.). Generating a reference image in such a manner may be performed in any suitable manner otherwise known in the art. In some such cases, a reference image may be created prior to imaging of the specimen and generation of test images used by the embodiments described herein. Thus, if such reference images are particularly noisy, they may be denoised as described herein prior to imaging of the specimen performed to generate test images. In this way, denoising of the reference image can be performed during the setup phase rather than the runtime phase if the reference image is not generated from the test specimen. However, even when a reference image is created using a test specimen, denoising of the reference image may be performed during the setup phase rather than the runtime phase.

In one embodiment, the reference image is a denoised reference image, the initial reference image is created for the specimen by the imaging subsystem at a location on the specimen that corresponds to the location at which the test image is created, and the one or more computer subsystems Reduce noise in the initial reference image prior to alignment to produce a denoised reference image, wherein reducing noise in the test image and reducing noise in the initial reference image are performed separately for the test image and the initial reference image. including structure-based denoising. One such embodiment is shown in FIG. 4 . As shown in FIG. 4 , test image 400 and reference image 402 may be separately fed to different structures based on denoising and ROI steps 404 and 406 , respectively. Both the structure-based denoising step and the ROI step may be performed as further described herein. The denoised test image and reference image with the identified ROI may then be input to an alignment step 408 . Alignment may be performed at step 408 as further described herein. Thus, in this relatively simple embodiment, the system performs structure-based denoising separately on these two images before feeding them to the main alignment step.

In a further embodiment, the reference image is a denoised reference image, the initial reference image is created for the specimen by the imaging subsystem at a location on the specimen that corresponds to the location at which the test image is created, and the one or more computer subsystems comprise: and to generate a denoised reference image by reducing noise in the initial reference image prior to alignment, wherein reducing noise in the test image and reducing noise in the initial reference image combine the test image and the initial reference image into a structure-based denoising image. Including simultaneous input into easing. Thus, in this embodiment, the computer subsystem(s) may create a “stack” from the test image and reference image prior to performing structure-based denoising on these images. For example, as shown in FIG. 5 , the computer subsystem(s) can create a test and reference stack 500 and input the image stack to a structure-based denoising step 502 . As used herein, the term “stacking” images does not include combining or altering images in any way. Instead, the term “stacked” as used herein is meant to indicate that multiple images are simultaneously fed to the denoising algorithm as different images of different input channels. This denoising step may be performed as further described herein. In this way, the computer subsystem(s) can use structure-based denoising on a stack of test images and reference images before separating them for the main alignment method.

After denoising, the images can be separated again and then fed to an alignment process. In one such embodiment, structure-based denoising simultaneously outputs a denoised test image and a denoised reference image, and one or more computer subsystems output the denoised test image and the denoised reference image prior to detection. configured to separate. For example, as shown in FIG. 5 , the output of structure-based denoising 502 is input to segmentation step 504 where denoised test image 506 is separated from denoised reference image 508. It can be. Then, ROI identification may be performed on the denoised test images and reference images 506 and 508 separately in ROI identification steps 510 and 512, respectively. The denoised test and reference images and their identified ROIs may then be input to an alignment step 514, which may be performed as further described herein.

The computer subsystem(s) may be configured to store the results of the alignment step to any suitable computer readable storage medium. This result may be stored along with any other result described herein and may be stored in any manner known in the art. The storage medium may include any storage medium described herein or any other suitable storage medium known in the art. After the results are stored, they can be accessed on a storage medium and used by any of the method or system embodiments described herein, formatted for display to a user, and used by another software module, method or system, or the like.

In one such example, an embodiment described herein may be configured to perform an inspection process on a specimen. In this way, test images and reference images (original and/or denoised) aligned with each other may be input into a defect detection method or algorithm by embodiments described herein. An example of such a defect detection method or algorithm may include subtracting a reference image from an aligned, denoised test image to create a difference image, and then applying a defect detection threshold to the difference image. For example, image signals or data of difference images that exceed a threshold may be designated as defects or latent defects, and image signals or data of difference images that do not exceed a threshold may not be designated as defects or latent defects. there is. Of course, this is merely an example of a defect detection method or algorithm that may use the results of the alignment steps described herein. In general, any defect detection method or algorithm that relies on alignment between a test image and a reference image can use the output of the alignment step described herein for inspection.

Different inspection tools may use different defect detection methods and algorithms. Some examples of defect detection algorithms used by commercially available inspection tools at KLA include performing a candidate image to reference image comparison by image frame subtraction, and a median reference frame of more than two frames. multi-die adaptive threshold (MDAT) algorithm, which identifies outliers based on signal-to-noise through double detection (comparing a candidate image to two reference images) or single detection when compared to include Another such defect detection algorithm includes the multi-computed die adaptive threshold (MCAT) algorithm, which is similar to the MDAT algorithm but optimizes the reference to be similar to the test image frame before image subtraction is performed. Additional such defect detection algorithms include the MCAT+ algorithm, which is similar to MCAT but uses references across the wafer. A further such defect detection algorithm is the single reference die (SRD) defect detection algorithm, which uses a reference die from the same or another wafer as a reference (to subtract from the test image).

The defect detection method may also be a one dimensional (1D) or two dimensional (2D) version of any of these defect detection methods. A defect detection method that generates a one-dimensional histogram for the detector output may be referred to as a “1D defect detection method”. In one such embodiment, a 1D histogram is generated from the gray levels of the difference image generated from the aligned image. For example, a 1D defect detection method or algorithm may use the 1D histogram for outlier detection along with the difference gray levels on the x-axis. Thus, a 1D histogram can show defect counts for different gray levels. In contrast, a "2D defect detection algorithm" as the term is used herein is, for example, one axis (y-axis) the central gray level of n>1 reference frames and the x-axis as described herein. It is an algorithm that uses a 2D histogram, which is the difference gray level of difference images generated from aligned images.

In another such example, embodiments described herein may be configured to perform a metrology process on a specimen. In this way, test images and reference images (original and/or denoised) aligned with each other can be input into measurement methods or algorithms that may be performed by embodiments described herein. Aligning the images with each other may be performed during metrology to determine relative characteristics, such as, for example, relative critical dimensions (CDs) of features in the test image compared to features in the reference image. In this way, it may be used to monitor the performance of processes performed on specimens for metrology purposes, may be used to identify problems in such manufacturing processes, may be used to determine calibrations for such manufacturing processes, or Differences between features of the test image and the reference image that can be used for , , and the like can be determined. These processes and corrections are further described herein.

In a further example, embodiments described herein may be configured to perform a defect review process on a specimen. In this way, test images and reference images (original and/or denoised) aligned with each other as described herein can be used to re-detect a defect detected in an inspection and possibly determine additional information about the defect. , which can be performed by the embodiments described herein. For example, if a defect is re-detected by subtracting a reference image from an aligned test image, the portion of the test image that corresponds to the defect has additional information about the defect as determined by inspection and/or a higher resolution than that determined by inspection. It can be used to determine the information determined by In one such example, because the defect review image is generated at a higher resolution than the inspection image, more accurate and/or detailed information may be determined by the defect review than the inspection. Since denoised test images can be used for this purpose, the defect information that can be determined from the test images described herein can also be more accurate and/or detailed. In some cases, it may be useful to determine one or more characteristics of a defect redetected by review using both the original and denoised test images.

Results and information generated from performing processes on specimens based on aligned images as described herein may be used in a variety of ways by the embodiments described herein and/or other systems and methods. These functions include, but are not limited to, process changes, such as, for example, manufacturing processes or steps that have been or will be performed on a specimen or another specimen in a feedback or feedforward manner. For example, the computer subsystem(s) may use the detected defect(s), measurements, reviewed defects, etc. to test specimens that have been inspected, measured, reviewed for defects, etc. as described herein. It may be configured to determine one or more changes to the process performed on and/or the process to be performed on the specimen. A change to a process may include any suitable change to one or more parameters of the process. The computer subsystem(s) can reduce or prevent defects on other specimens where the modified process is performed, and defects and/or measurements can be corrected on the specimen in another process performed on the specimen, and Desirably determines such changes so that in another process performed on the system, defects and/or measurements can be compensated for, etc. The computer subsystem(s) may determine such a change in any suitable manner known in the art.

These changes can then be transmitted to a semiconductor manufacturing system (not shown) or a storage medium (not shown) accessible to the computer subsystem(s) and the semiconductor manufacturing system. A semiconductor manufacturing system may or may not be part of a system embodiment described herein. For example, other computer subsystem(s) and/or imaging subsystems described herein are coupled to the semiconductor manufacturing system through one or more common elements, such as, for example, housings, power supplies, specimen handling devices or mechanisms, and the like. It can be. The semiconductor manufacturing system may include any semiconductor manufacturing system known in the art, such as, for example, lithography tools, etching tools, chemical-mechanical polishing (CMP) tools, deposition tools, and the like.

The embodiments described herein provide many advantages over previously used methods and systems for image alignment. For example, the embodiments described herein enable image alignment on relatively noisy images, such as SEM images, which previously had relatively poor performance using previous alignment methods. Achieving better image alignment on noisy images is possible using structure-based denoising and/or SVD described here. In another example, embodiments described herein may close gaps in current image alignment, eg, current SEM image alignment, when the test image and the reference image have less obvious features in common and/or have noise. can Embodiments described herein detect horizontal features and/or vertical features and set these features in ROIs as inputs to an alignment step, which may include NCC or some other suitable alignment method or algorithm, in a relatively limited number of ways. It can outperform previous alignment methods for images with a structure of

Each of the embodiments of each of the systems can be combined together into one single embodiment.

Another embodiment relates to a computer-implemented method for aligning an image of a specimen. The method includes the noise reduction step, one or more patterned feature detection steps, region designation steps, and alignment steps, as described above.

Each step of this method may be performed as further described herein. The method may also include any other step(s) that may be performed by the imaging subsystem and/or computer subsystem(s) or system(s) described herein. Noise reduction steps, one or more patterned feature detection steps, region designation steps, and alignment steps are performed by one or more computer subsystems coupled to the imaging subsystem, all of which may be implemented in any of the embodiments described herein. can be configured accordingly. Additionally, the methods described above may be performed by any of the system embodiments described herein.

A further embodiment relates to a non-transitory computer readable medium storing program instructions executable on a computer system to perform a computer implemented method for aligning an image of a specimen. One such embodiment is shown in FIG. 7 . In particular, as shown in FIG. 7 , non-transitory computer readable medium 700 includes program instructions 702 executable on computer system 704 . A computer-implemented method may include any step(s) of any method(s) described herein.

Program instructions 702 implementing methods such as those described herein may be stored on computer readable medium 700 . A computer readable medium may be, for example, a storage medium such as a magnetic disk or optical disk, magnetic tape, or any other suitable non-transitory computer readable medium known in the art.

Program instructions may be implemented in any of a variety of ways, including, among others, procedure-based, component-based, and/or object-oriented technologies. For example, program instructions may be implemented using ActiveX controls, C++ objects, JavaBeans, Microsoft Foundation Classes ("MFC"), Streaming SIMD Extensions (SSE), or other technologies or methodologies, as desired. there is.

Computer system 704 may be configured according to any of the embodiments described herein.

Additional modifications and alternative embodiments of the various aspects of the present invention will become apparent to those skilled in the art in light of this description. For example, a method and system for aligning an image to a specimen is provided. Accordingly, these descriptions are to be construed as illustrative only, and are intended to teach one of ordinary skill in the art the general way of carrying out the present invention. It should be understood that the forms of the invention shown and described herein are to be taken as presently preferred embodiments. As will be apparent to those skilled in the art after all having the benefit of this description of the invention, elements and materials may be substituted for those shown and described herein, and parts and processes may be reversed. and certain features of the present invention may be used independently. Changes may be made in the elements described herein without departing from the spirit and scope of the invention as set forth in the claims below.

Claims (20)

In a system configured to align an image of a specimen,
an imaging subsystem configured to generate an image of the specimen; and
one or more computer subsystems
Including, the one or more computer subsystems,
reduce noise in a test image generated for the specimen by the imaging subsystem to produce a denoised test image;
detect one or more patterned features in the denoised test image that extend in at least a horizontal or vertical direction;
designate a region of the denoised test image in which the detected one or more patterned features are located as a region of interest in the denoised test image;
Align the denoised test image to a reference image for the specimen using only the region of interest in the denoised test image and a corresponding region in the reference image.
A system configured to align an image of a specimen, wherein the system is configured. 2. The method of claim 1 , wherein the reference image is a denoised reference image, an initial reference image is created for the specimen by the imaging subsystem at a location on the specimen corresponding to a location at which the test image is created, and wherein the one or more computer subsystems are also configured to reduce noise in the initial reference image prior to the alignment to produce the denoised reference image. The system of claim 1 , wherein the reference image is not created using the specimen. 2. The method of claim 1 , wherein the reference image is a denoised reference image, an initial reference image was not created using the specimen, and the one or more computer subsystems further reduce noise in the initial reference image prior to the alignment. reduction to generate the denoised reference image. 2. The system of claim 1, wherein reducing noise comprises structure based denoising. 2. The method of claim 1 , wherein the reference image is a denoised reference image, an initial reference image is created for the specimen by the imaging subsystem at a location on the specimen corresponding to a location at which the test image is created, and The one or more computer subsystems are also configured to generate the denoised reference image by reducing noise in the initial reference image prior to the alignment, reducing noise in the test image and noise in the initial reference image. Wherein reducing ? comprises structure-based denoising performed on the test image and the initial reference image separately. 2. The method of claim 1 , wherein the reference image is a denoised reference image, an initial reference image is created for the specimen by the imaging subsystem at a location on the specimen corresponding to a location at which the test image is created, and The one or more computer subsystems are also configured to generate the denoised reference image by reducing noise in the initial reference image prior to the alignment, reducing noise in the test image and noise in the initial reference image. Wherein reducing ? comprises simultaneously inputting the test image and the initial reference image to structure-based denoising. 8. The method of claim 7, wherein the structure-based denoising simultaneously outputs the denoised test image and the denoised reference image, and wherein the one or more computer subsystems also: prior to the detection, the denoised test image and the denoised test image. And configured to isolate the denoised reference image. The system of claim 1 , wherein reducing the noise comprises singular value decomposition. 2. The method of claim 1, wherein reducing the noise comprises singular value decomposition, selecting a predetermined number of eigenvectors having a maximum value in one or more matrices output by the singular value decomposition, and performing the test from the selected eigenvectors. and reconstructing an image to generate the denoised test image. 2. The system of claim 1, wherein reducing noise comprises structure-based denoising followed by singular value decomposition. The system of claim 1 , wherein aligning the denoised test image to a reference image comprises aligning the test image to the reference image with sub-pixel accuracy. 2. The system of claim 1, wherein the sorting comprises normalized cross-correlation. 2. The system of claim 1, wherein a majority of patterned features extending in at least the horizontal direction or the vertical direction in the test image have a smaller weight in normalized cross correlation than noise in the test image. 2. The system of claim 1, wherein the test image includes repetitive patterned features. The system of claim 1 , wherein the specimen is a wafer. The system of claim 1 , wherein the imaging subsystem is a light-based imaging subsystem. The system of claim 1 , wherein the imaging subsystem is an electronic based imaging subsystem. A non-transitory computer-readable medium storing program instructions executable on a computer system to perform a computer-implemented method for aligning an image of a specimen, the computer-implemented method comprising:
reducing noise in a test image generated for the specimen by the imaging subsystem to produce a denoised test image;
detecting one or more patterned features in the denoised test image that extend in at least a horizontal or vertical direction;
designating an area of the denoised test image in which the detected one or more patterned features are located as a region of interest in the denoised test image; and
aligning the denoised test image to a reference image for the specimen using only the region of interest in the denoised test image and a corresponding region in the reference image - the reducing, the detecting wherein the designating, and the aligning are performed by the computer system coupled to the imaging subsystem -
A non-transitory computer readable medium comprising a. In the computer-implemented method for aligning the image of the specimen,
reducing noise in a test image generated for the specimen by the imaging subsystem to produce a denoised test image;
detecting one or more patterned features in the denoised test image that extend in at least a horizontal or vertical direction;
designating an area of the denoised test image in which the detected one or more patterned features are located as a region of interest in the denoised test image; and
aligning the denoised test image to a reference image for the specimen using only the region of interest in the denoised test image and a corresponding region in the reference image - the reducing, the detecting wherein the specifying, and the aligning are performed by one or more computer subsystems coupled to the imaging subsystem;
A computer-implemented method for aligning an image of a specimen, including a.

Images